• 23 September 2026
• Updated: 25 September 2026
• Mr.Pendragon

Market bugs top the pain chart. Storage tops the wish list

SEED's market is getting substantially safer, but it has not earned blind trust. On 11 September, Klang reported that market failures had fallen from around 20% to around 4%, with the biggest improvements affecting faults that lost items or money and left orders blocked. That figure is a snapshot, not a fresh measurement after the subsequent fixes.

That was the good news in Klang's original update. The studio has since detailed further market and Auto-Restock fixes in its 0.3.5 patch notes, on 21 September it closed a small group of legacy cross-society sell orders, and on 24 September it refunded 21.2 million SeedCoin for three older market faults. A delivery can still be held up by a full inbox, no Seedling scheduled to make the trip, or a carrier who falls unconscious while transporting your purchase. Avesta has managed to make checkout a staffing decision.

This analysis is based on Klang's 11 September 2026 market update, its earlier market report, the 0.3.5 patch notes, shipped game data a 21 September message from Klang's Andy and a 24 September message from Klang's Fito, both in the official SEED Discord. The percentages and order counts are Klang's figures, not AnvilGaming test results.

24 September update: 21.2 million SeedCoin refunded

Klang has paid out compensation for three market faults. Klang's Fito posted in the official SEED Discord on 24 September that the studio had credited 21.2 million SeedCoin to more than 890 players, across 1,244 wallets: 507 Seedling wallets, 656 business wallets, 65 group wallets and 16 society treasuries. The refund followed a promise made on Klang's Notice Board stream the day before.

• Cross-society orders that were partially delivered but charged in full, 6 to 19 August.
• Import tax charged at 100 times the rate, 27 to 29 July.
• Goods destroyed because they could not be received at your Spaceport, 25 August to 16 September.

If you were affected, the credit is already in the wallet that paid for the order, labelled "Issue Compensation" in the wallet history. Klang says this won't cover everyone, because some losses leave no trace it can find; if you lost SeedCoin or goods to a market bug and see no credit, contact support with the order details. Klang has also removed the old, inactive buy orders that showed as "N/A" in the market.

That is compensation for past faults, not a new failure rate. Klang has still not published a figure to replace the 4% from 11 September.

21 September update: legacy sell orders closed

Klang's Andy says engineers have closed the remaining legacy cross-society sell orders that used the old Spaceport-first system. Since early August, new orders have listed directly from a seller building's Outbox instead. The old method could leave unsold goods sitting at a Spaceport, with some items stuck there.

If one of your older orders has disappeared, check the Outbox and other containers in the seller building before re-listing the goods at a time that suits you. Klang says orders placed since the early-August change should be unaffected.

Klang says the clean-up affects 355 sell orders out of roughly 18,000 that were open at the time. That is a count of legacy orders closed, not a new market-failure rate, and we have not independently verified the return of individual players' goods.

From one order in five to roughly one in twenty

Klang's 11 September market reliability update put the previous failure rate at around 20% and the rate measured then at around 4%. Taken at face value, that is a reduction of roughly four-fifths.

More important than the headline number is where the improvement happened. Klang says the largest gains are in failures that could lose items or money, or leave an order blocked. Those were the faults that made trade dangerous rather than merely irritating.

The number still needs context. Klang does not define the measurement window, whether the percentage counts orders or delivery attempts, which markets are included or whether legitimate simulation failures sit in the same figure as software faults. That does not erase the improvement. It does mean nobody outside Klang can reproduce the calculation or treat 4% as a warranty.

As of our 11 September check, we had not encountered a market failure since 0.3.1. That was genuinely encouraging. Because we did not document the number and type of orders behind that experience, however, it remains a direct observation rather than a competing failure rate.

Klang did not call the job finished. Its 11 September post named Auto-Restock as one remaining problem; the 0.3.5 patch notes, published two days later, list several fixes to automatic restocking and other market behaviour. They do not supply a revised failure rate, so the 4% figure should not be read as a post-0.3.5 result.

The market is also a logistics system

One useful part of Klang's 11 September explanation is that many failed deliveries were being caused by the simulation around the market.

A buyer needs room in the destination inbox or storage. A self-pickup needs an available Seedling scheduled to Haul. A logistics company needs employees assigned to Client Deliveries. The delivery can still be interrupted after collection if the carrying Seedling keeps the item, falls unconscious or dies and drops it somewhere.

The shipped delivery activity supports that explanation. Before a haul can proceed, the game checks that the assignment permits delivery and that the destination has capacity. The work itself is not an abstract transfer: a Seedling walks to the source, picks the resource up, carries it to the destination and puts it down.

That distinction matters. A market order can match correctly and still fail because the physical work needed to finish it never happens. Fixing the market code will not create storage space or put a logistics employee on shift.

It is also a difficult line to draw fairly when the game does not explain the failure. Klang acknowledges that problem and says the interface and notifications need to tell players what happened and what they can do about it. Until those messages are reliable, a legitimate simulation outcome and a bug can look exactly alike from the wrong side of a missing shipment.

Who carries the loss changes during the journey?

Klang has now explained how responsibility is assigned when an item is genuinely lost through the simulation rather than through a bug.

• Same-sector delivery: the buyer becomes responsible once the goods have been picked up from the seller's outbox.
• Cross-sector delivery: the seller remains responsible until the goods reach the seller's Spaceport. Responsibility then passes to the buyer.
• Bug-caused loss: Klang says this is not a legitimate simulation outcome. The studio investigates whether compensation is appropriate, and any compensation should come from the game rather than another player.

The principle is sensible: responsibility follows the part of the route a player controls. It also explains a refund change. Under the older behaviour, a buyer who lost an item during delivery could be refunded from the seller's account even after the seller had completed their part. That punished the wrong player and left a route for exploitation.

There is another complication. An interrupted hauler may still have the item or may have dropped it in the world. Automatically creating compensation before establishing what happened risks duplicating the goods. A missing delivery is therefore not automatically a destroyed delivery.

If an order goes wrong, trace the carrier and the route before assuming the stock has vanished. It may be sitting in a Seedling's inventory while everyone involved is having a substantially worse day than the order screen suggests.

Old and new Delivery Inboxes behave differently

One practical difference can now be confirmed: a Delivery Inbox placed before 0.3.4 can retain its old request settings, while one added after the update begins with nothing selected.

The 11 September update warns that an inbox accepting every kind of item can be filled by ordinary hauling. Klang's 0.3.4 patch notes, published four days earlier, say a newly placed Delivery Inbox starts by requesting nothing instead of everything. The shipped container definition agrees with the 0.3.4 notes. Those descriptions appear contradictory until the age of the container is taken into account.

We have now seen both behaviours in Wyrmhold. The existing Delivery Inbox serving our rental properties still has every resource selected. The Inbox in a residence built for our Seedlings after 0.3.4 defaulted to nothing selected. That live comparison supports the old-versus-new explanation.

Our 0.3.4 analysis reported the new default described in those patch notes. The practical rule is to inspect the actual Inbox. An older one may fill with unrelated stock; a newer one may reject everything until its request list is configured. Before placing a market order, confirm that the destination has free capacity and requests the item you are buying.

The wider direction is clearer too. Klang's September figures showed fewer market failures, our own use had been clean since 0.3.1 when we checked on 11 September, and the studio has since documented further fixes and this legacy-order clean-up.

The market is moving in the right direction

For sellers caught by the old Spaceport-first system, the immediate step is to look in the original building and re-list any returned goods when ready. For everyone else, the 21 September clean-up is welcome progress on a specific problem, not proof that every order now succeeds. Klang has not published a new failure rate, and the capacity and delivery-labour checks still matter.